Free tools Windows power users keep installed
One-click scans. No signup required.
AI drug discovery is not a single kind of product, and these nine examples are not a ranked list. They range from molecular-design software and systems linking machine learning to automated experiments to shared pharmaceutical models and cloud collaborations. The useful distinction is what each system does, how a pharma partner accesses it, and what evidence has actually been disclosed—not whether a company uses the word “AI.”
What counts as an AI drug discovery platform?
The label covers tools with different jobs in research and development. Some predict protein structures or help design molecules; others connect experimental data, automated laboratory work and machine learning. Still others make a pharmaceutical company’s predictive models available to selected collaborators or supply cloud and foundation-model services within an R&D partnership. They are not interchangeable products, and the partnerships described here do not establish that every system is available to buy or use independently.
As an Amazon Associate I earn from qualifying purchases.
The nine examples below are a curated set supported by public company and partnership announcements, not a canonical market list or an independent performance ranking. Announcements establish that a collaboration or stated aim exists; they do not, on their own, show that AI shortened discovery or improved clinical success.
Which AI drug discovery platforms have pharma activity?
| Example | Operating model | Disclosed pharma activity |
|---|---|---|
| AWS AI for Novo Nordisk | Cloud and AI collaboration | Novo Nordisk named AWS its preferred cloud provider and strategic AI partner in August 2026. |
| Iambic Therapeutics | Computational small-molecule discovery | Bayer announced a collaboration in June 2026 focused on hard-to-drug targets. |
| Exscientia | End-to-end small-molecule discovery and translational research | Sanofi describes a collaboration targeting cancer and immune-mediated diseases. |
| BioMap | AI modules and protein language models for biologics | Sanofi describes co-development for biologics design and optimization. |
| Recursion OS | Integrated machine learning and experimental discovery system | Sanofi describes a partnership launched in 2022, with multiple programs reaching development milestones. |
| Schrödinger computational platform | Computational chemistry and molecular modeling software | Schrödinger reports software licensing; its 2026 BMS announcement concerns Bunsen AI co-scientist deployment. |
| Lilly TuneLab via Revvity Signals Xynthetica | Collaborative access to predictive models | Revvity announced in January 2026 that Lilly models would be available through its Signals platform to selected participants. |
| Isomorphic Labs Drug Design Engine (IsoDDE) | Predictive and generative AI for biology and molecule design | Isomorphic’s May 2026 financing announcement identifies continued engine development and deployment. |
| Insilico Medicine Pharma.AI | End-to-end AI drug-discovery offering | A December 2025 company filing describes collaborations with 13 of the 20 largest pharmaceutical companies by reported 2024 sales. |
AWS AI for Novo Nordisk
Novo Nordisk’s August 2026 announcement calls AWS its preferred cloud provider and strategic AI partner, and describes a co-innovation hub in London. Named services include Amazon Bio Discovery and Amazon Bedrock, intended to support target identification, therapy design and connections among genomic, imaging and clinical data. This is a cloud-and-AI collaboration, not evidence that AWS by itself supplies a complete drug-discovery platform. The announcement also reports productivity results in other areas, such as clinical documentation time and employee enablement; those are not drug-discovery outcomes.
#1 Best Overall
Iambic Therapeutics: Enchant and NeuralPLexer
Bayer’s June 2026 collaboration with Iambic focuses on small-molecule discovery for hard-to-drug targets. Bayer names Iambic’s Enchant and NeuralPLexer technologies and says the work aims to find novel entry points and differentiated molecules. Iambic describes Enchant as a multimodal transformer and NeuralPLexer as a protein–ligand structure-prediction technology. The announcement establishes the partnership and its aims, not a resulting clinical benefit.
Exscientia
Sanofi describes Exscientia as part of an end-to-end AI platform for drug discovery and translational research in cancer and immune-mediated diseases. Sanofi sets an ambition of generating up to 15 small-molecule development candidates through the collaboration. “Up to 15” is a target, not a count of candidates already delivered.
Rank #2
BioMap
BioMap provides a biologics-focused contrast to the small-molecule examples. Sanofi says the companies are co-developing AI modules and protein language models for biologics design and multiparametric optimization. The public description states what the work is intended to do; it does not establish completed product validation.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Recursion OS
Sanofi’s 2026 spotlight describes its partnership with Recursion as launched in 2022, covering small-molecule programs in immunology and oncology. Sanofi says multiple programs advanced and reached development milestones. Recursion presents Recursion OS as a system spanning target identification through clinical-trial enrollment, combining wet-lab automation, data and machine learning. That integrated experimental-and-computational loop is the key distinction from software focused primarily on molecular modeling. Company descriptions of scale or speed should not be read as independently verified comparisons.
Schrödinger’s computational platform
Schrödinger’s platform is computational chemistry and molecular modeling infrastructure for molecular discovery and optimization, distinct from companies that operate their own high-throughput wet labs. Schrödinger says it has invested in the platform for more than 30 years and licenses its software to industry and academic users. Its 2026 announcement with Bristol Myers Squibb concerns deployment of Bunsen, an AI co-scientist for agentic discovery; it is an announced collaboration, not evidence of a delivered therapeutic outcome.
Lilly TuneLab through Revvity Signals Xynthetica
Revvity’s January 2026 announcement says Lilly predictive models trained on Lilly research data are available through the Signals platform. The described federated-learning framework lets participants contribute data and use models while keeping proprietary data private. This is collaborative model access, not necessarily a standalone product that any organization can purchase. Lilly and Revvity said they would jointly fund access for selected participants, so access should not be treated as universally open.
Isomorphic Labs Drug Design Engine
Isomorphic describes IsoDDE as an engine using predictive and generative AI models for biological phenomena and molecule design. Its May 2026 financing announcement identifies continued development and deployment of the engine. That supports including it as a drug-design platform example, but does not establish that it is superior to competing systems or has produced therapeutic success.
Insilico Medicine Pharma.AI
A December 2025 company filing describes Pharma.AI as an end-to-end offering spanning target identification, small-molecule generation and clinical-outcome prediction. The filing reports collaborations with 13 of the 20 largest pharmaceutical companies by reported 2024 sales. That company-reported count does not reveal the depth, scope or current status of each relationship.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How should pharma teams compare these platforms?
A meaningful comparison starts with the research task, not a vendor’s broad claim to cover drug discovery. A group choosing among systems should establish what stage it needs help with and then ask how the technology fits existing experiments, data policies and decision-making.
- Scope and modality: Is the work about target identification, small-molecule design, biologics, optimization, or another stage? A protein language model and a molecular-modeling package address different problems.
- Wet-lab integration: Does the system operate experiments and feed results back into its models, as Recursion describes, or primarily provide computational tools?
- Data access and governance: Find out what data must be shared, where it is stored, who can use it, and whether the collaboration uses federated access or another privacy arrangement.
- Access model: Distinguish licensed software, a co-development alliance, cloud services, and invitation-based access to a partner’s models. A named partnership does not prove that the platform is offered to other buyers.
- Evidence and maturity: Separate a stated research goal from a completed milestone, published validation and a clinical outcome. Ask what evidence applies to the specific program and task being evaluated.
Can AI make drug discovery faster?
AI can be used to prioritize targets, model molecular interactions, generate candidate molecules and connect experimental results to the next design cycle. Those are plausible ways to assist research workflows, but the public announcements and company descriptions summarized here do not provide a comparable, independently verified statistic for industry-wide gains in discovery time, cost or clinical success attributable to AI. A productivity result in one business process, a company’s platform claim or a partnership announcement cannot substitute for that evidence.
One alternative with a narrower focus
AlgoraeOS is a possible alternative if a shortlist specifically needs a platform centered on drug-combination candidates. Algorae says it integrates preclinical, clinical, chemical and biological data to generate candidates for licensing and co-development. The company reports training on more than 5.5 million unique inhibition records and testing 21 drug-drug targets across four cancer cell lines. Those are company-reported figures, and one listed publication is marked pending; they are not a substitute for independent peer-reviewed assessment.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
